harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexei Fedotov" <alexei.fedo...@gmail.com>
Subject Re: [tests] cpd report available
Date Wed, 22 Nov 2006 22:12:55 GMT
I like it!

When I ran CPD on C code base it discovered more than cut&paste
problems. Usually the code which contained big cut&pasted fragments
had design problems as well. I wonder if it is true for Harmony.

1. I have no idea if and how java.* packages should be fixed. We don't
have much freedom there.
2. The first comment of X11 and Win32 files from nativebridge package
tells that they are autogenerated. Should we do anything about them?

I'm eager for DRLVM and test results.

-- 
Thank you for a wonderful page,
Alexei

On 11/22/06, Stefano Mazzocchi <stefano@apache.org> wrote:
> Paulex Yang wrote:
> > Stefano Mazzocchi wrote:
> >> I've managed to run CPD[1] against harmony classlib.
> >>
> >> CPD (or cut/paste detector) is a tool that is capable of showing where
> >> code was cut/pasted from different parts of the source tree.
> >>
> >> I managed to filter out all the tests (because they seem to have a good
> >> amount of cut/paste and that is not necessarily a problem) but there is
> >> still a lot of code that is was cut/pasted and might be worth
> >> considering for refactoring.
> >> Find it at
> >>
> >>  http://67.86.14.213:10000/results/latest/cpd.html
> >>
> > Cool! possible to exclude some special classes, such as Local_***.java
> > and nls/Messages.java? They are either only data, or intentionally keep
> > a copy in every module.
>
> sure thing. The data is dumped by CPD in XML and then I use an XSLT
> stylesheet to generate that page, I can filter out whatever we want.
>
> I'll remove Locals and Messages and regenerate.
>
> --
> Stefano.
>
>

Mime
View raw message